Báo cáo khoa học: Structural basis for recognition of Co2+ by RNA aptamers pot

Structural basis for recognition of Co
2+
by RNA aptamers
Jan Wrzesinski and Stanisław K. Jo
´
z
´
wiakowski
Institute of Bioorganic Chemistry, Polish Academy of Sciences, Poznan
´
, Poland
RNA molecules are involved in numerous fundamental
cellular processes, such as replication, transcription
and translation. Most recently, participation of small
interfering RNA, microRNA and noncoding RNA in
the regulation of gene expression and the development
of variety organisms has been intensively investigated
[1–3]. However, the interaction of RNA molecules with
cellular components requires their proper folding into
the active structure. This process is facilitated by the
presence of cations, such as polyamines and mono-
valent and divalent metal ions [4,5]. Determination of
the precise location of the metal ions inside the RNA
structure is important for a better understanding of
RNA interactions with other components of the cell.
Hence, many biophysical and biochemical methods
have been developed for defining RNA ligands that
coordinate metal ions.
The most informative biophysical methods, which
involve X-ray crystallography and NMR spectroscopy,
provide details of the structure of the metal-ion bind-

ing sites and their coordination spheres [6,7]. The
disadvantages of X-ray crystallography and NMR
spectroscopy include problems with respect to crystalli-
zation and the need for isotope enrichment to resolve
the RNA spectrum. Therefore, in order to gain a glo-
bal insight into metal ion–RNA interactions, it is often
necessary to conduct structural studies of metal ion
binding modes in RNA molecules using biophysical
methods simultaneously with other biochemical
approaches.
Metal ion-induced cleavage, an alternative approach
of biochemical studies, is frequently used to identify
those RNA stretches involved in the organization of
metal ion binding site(s) in solution. This approach
Keywords
Co
2+
binding RNA aptamers; Co
2+
-induced
conformational changes; kissing dimer;
NAIM; oligomer hybridization ⁄ RNase H
digestion
Correspondence
J. Wrzesinski, Institute of Bioorganic
Chemistry, Polish Academy of Sciences,
Noskowskiego 12 ⁄ 14, 61-704 Poznan
´
,
Poland

Fax: +48 61 8520532
Tel: +48 61 8528503
E-mail:
(Received 8 October 2007, revised 27
December 2007, accepted 5 February 2008)
doi:10.1111/j.1742-4658.2008.06320.x
Co
2+
binding RNA aptamers were chosen as research models to reveal
the structural basis underlying the recognition of Co
2+
by RNA, with the
application of two distinct methods. Using the nucleotide analog interfer-
ence mapping assay, we found strong interference effects after incorpora-
tion of the 7-deaza guanosine phosphorotioate analog into the RNA chain
at equivalent positions G27 and G28 in aptamer no. 18 and G25 and G26
in aptamer no. 20. The results obtained by nucleotide analog interference
mapping suggest that these guanine bases are crucial for the creation of
Co
2+
binding sites and that they appear to be involved in the coordination
of the ion to the exposed N7 atom of the tandem guanines. Additionally,
most 7-deaza guanosine phosphorotioate and 7-deaza adenosine phos-
phorotioate interferences were located in the common motifs: loop E-like
in aptamer no. 18 and kissing dimer in aptamer no. 20. We also found that
purine rich stretches containing guanines with the highest interference val-
ues were the targets for hybridization of 6-mers, which are members of the
semi-random oligodeoxyribonucleotide library in both aptamers. It tran-
spired that DNA oligomer directed RNase H digestions are sensitive to
Co

2+
and, at an elevated metal ion concentration, the hybridization of
oligomers to aptamer targets is inhibited, probably due to higher stability
and complexity of the RNA structure.
Abbreviations
c
7
AaS, 7-deaza adenosine phosphorotioate; c
7
GaS, 7-deaza guanosine phosphorotioate; NAIM, nucleotide analog interference mapping;
NTA, nitrilotriacetate.
FEBS Journal 275 (2008) 1651–1662 ª 2008 The Authors Journal compilation ª 2008 FEBS 1651
has been successfully used to detect various metal ions
(e.g. Mg
2+
,Pb
2+
,Mn
2+
,Eu
2+
,Tb
3+
,Co
2+
, etc.) in
divergent RNA molecules involving tRNAs, ribo-
zymes, and other RNA molecules [8–10].
The active involvement of metal ions in the cleavage
of several ribozymes has been investigated using a

‘metal ion specificity switch assay’ with respect to sul-
fur substitution of the oxygen atom in the phosphate
group [11]. The cleavage efficiency for such modified
ribozymes, strongly reduced in the presence of Mg
2+
,
is rescued when thiophillic metal ions Mn
2+
,Cd
2+
or
Zn
2+
are added [12,13].
However, the precise determination of RNA ligands
necessary for metal ion binding and RNA activities
(mainly ribozymes) has become possible by applying
nucleotide analog interference mapping (NAIM).
In vitro transcription in the presence of 5¢-O-(1-thio)-
nucleoside triphosphates enables incorporation of these
modified nucleotides or nucleoside analogs with altered
base moieties into the RNA chain, and I
2
cleavage
allows determination of the atom in the bases that
interferes with function [14,15]. NAIM has been used
to identify the RNA ligands that interact with metal
ions and the nucleotide modifications that are critical
for retaining ribozyme activity [16–18].
In the present study, we applied the NAIM method

to study the structural basis of the molecular mecha-
nism underlying the binding of Co
2+
to RNA mole-
cules using two aptamers, no. 18 and no. 20, which
coordinate Co
2+
ions, previously selected in our labo-
ratory, as research models [19]. Several purine N7
groups that interfere with Co
2+
were detected.
Additionally, the influence of the Co
2+
binding
on the aptamer structure using hybridization of a
6-mers semi-random oligodeoxynucleotide library and
RNase H digestion was investigated.
Results and Discussion
Synthesis of phosphorothioate nucleoside
modified RNA aptamers
As prepared using the T7 transcription system, a pool
of aptamers (Fig. 1) carrying randomly distributed
phosphorothioate modifications, as well as 7-deaza
guanine or adenine analog substitutions, was analyzed
to determine which RNA ligands interfere with Co
2+
binding. It is worth noting that T7 RNA polymerase
only incorporates the S
P

-stereoisomer of phosphoro-
thioate modified nucleotides into the RNA chain and
causes inversion of the configuration at the a-phospho-
rus atom, resulting in R
P
-phosphorothioate substitu-
tion [16]. We used nitrilotriacetate (NTA) resin with
immobilized Co
2+
to separate the partially modified
aptamers into Co
2+
binding and nonbinding fractions
by the metal ion affinity chromatography approach
and fractions were then subjected to iodine cleavage. A
comparison of cleavage patterns of both fractions
enables determination of which RNA ligands actively
participate in the Co
2+
binding event.
Effect of R
P
-phosphorothioate nucleoside
modification on Co
2+
binding
Only three phosphorothioate interferences within the
loop region of both aptamers were observed (Figs 2–
4). Two weak UaS(j ¼ 1.8) interferences took place
at U42 in aptamer no. 18 and at U44 in aptamer

no. 20 (in equal positions; the second nucleotide at the
3¢-end of the loop; Fig. 4). The third CaS moderate
interference (j ¼ 2.4) occurred at C36 in aptamer
no. 18.
Generally, relatively low interference effects were
discovered after replacing the nonbridging pro-R
P
oxy-
gen atom with sulfur. The reason for the above obser-
vation might be the dominant contribution of base
Fig. 1. The secondary structure of the in vitro selected aptamers
that bind Co
2+
. The locations of Co
2+
-induced cleavage sites are
shown by open arrows.
Recognition of Co
2+
by RNA aptamers J. Wrzesinski and S. K. Jo
´
z
´
wiakowski
1652 FEBS Journal 275 (2008) 1651–1662 ª 2008 The Authors Journal compilation ª 2008 FEBS
moieties in the formation of the Co
2+
binding site
architecture in the aptamers. Therefore, replacement of
the nonbridging oxygen atom with sulfur in the phos-

phate group had a small effect on Co
2+
binding in
contrast to ‘soft’ thiophillic metal ions Cd
2+
,Mn
2+
and Zn
2+
, which discriminated between the oxygen
and sulfur atoms, with the latter being preferentially
bound [20].
Importance of the N7 group of purines for the
development of Co
2+
binding sites
In aptamer no. 18, we found weak 7-deaza adenosine
phosphorotioate (c
7
AaS) interferences at A26
(j = 1.9) and A41 (j = 1.4) (Figs 2 and 4). By con-
trast, for 7-deaza guanosine phosphorotioate (c
7
GaS),
five interferences were observed and three of them
were localized at the 5¢-side of the loop. The strongest
interferences (j = 4.0 and 3.6) were detected at G27
and G28, respectively. Interference at G25 was less
prominent (j = 1.9). Additionally, there were two
weak interferences at G37 (j = 2.0) and G38

(j = 1.9), within the previously identified Co
2+
-
induced cleavage sites [19]. The strongest interferences
occurred at G27 and G28, which are positioned within
the purine rich stretch 19-AGGCGAGAGG-28. It is
known that such regions containing purine stretches
are often involved in strong stacking interactions,
reducing their flexibility [21,22]. Thus, positioned
within the stacked, more rigid 19-AGGCGAGAGG-28
stretch, guanine bases would be accessible to interact
with the Co
2+
ion immobilized on NTA resin, and the
ion is probably coordinated to N7 atom of the imidaz-
ole ring of guanines. Yet the possibility that deletion
of the N7 group of purines may result in an alternative
structure of the aptamer cannot be excluded. However,
Fig. 2. Iodine cleavage analysis of phospho-
tioate nucleoside analogs modified 5¢-
32
P
labeled Co
2+
binding aptamer no. 18. Lanes:
C, reaction control of the F
N
fraction; F
N
,

RNA fraction which is not bound to Co
2
-
NTA resin; F
B
, RNA fraction that is effec-
tively bound to Co
2+
-NTA resin and is eluted
with 2 m
M concentration of Co
2+
; L, form-
amide ladder; T1, limited hydrolysis by
RNase T
1
. Guanine residues are labeled on
the right. Sites of interference are denoted
with arrows.
J. Wrzesinski and S. K. Jo
´
z
´
wiakowski Recognition of Co
2+
by RNA aptamers
FEBS Journal 275 (2008) 1651–1662 ª 2008 The Authors Journal compilation ª 2008 FEBS 1653
these purines are located in the single-stranded region
accessible for chemical probing [19]. Thus, N7 atoms
of the above guanine residues are not involved in inter-

actions with other base ligands or in aptamer structure
development. Interestingly, similar binding properties
of Co
2+
to the N7 atom of guanine residues were
previously observed in the crystal structure of the
orthorombic form of yeast tRNA
Phe
[23].
In the predicted loop E-like motif, bases G25 ⁄ A43,
as well as G27 ⁄ A41, form a sheared G-A base pairing,
whereas the A26-U42 base pair displays a reversed
Hoogsteen geometry [24] (Fig. 5). Thus, c
7
AaS interfer-
ences at A26 and A41 confirm that the N7 atom of
these adenine bases is involved in the formation of non-
standard H-bonding. Additionally, in the proposed
loop E-like motif structure, the aforementioned N7
position of G25 is exposed to interaction with Co
2+
and such interactions may stabilize this motif.
Furthermore, the N7 atom of G37 and G38 partici-
pates in Co
2+
coordination or is involved in other
interactions that build Co
2+
binding sites. However,
the interference j values are at least two-fold lower in

comparison with G27 and G28, indicating that those
interactions are weaker.
Unlike in aptamer no. 18, the c
7
AaS interferences
were not observed in aptamer no. 20 (Figs 3 and 4). In
the case of the c
7
GaS modified aptamer, four inter-
ferences were found. Strong interferences at G25
(j = 3.7) and G26 (j = 3.5) are located at the same
positions as the interferences at G27 and G28 in apt-
amer no. 18 (i.e. in the third and fourth positions at
the 5¢-end of loop, within 19AGGCGA
GG-26, a pur-
ine stretch two nucleotides shorter than in the case of
aptamer no. 18). The determined j values were very
similar: 4.0 and 3.6 for G27 and G28 in aptamer
no. 18 and 3.7 and 3.5 for G25 and G26 in aptamer
Fig. 3. Iodine cleavage analysis of phospho-
tioate nucleoside analogs modified 5¢-
32
P
labeled Co
2+
binding aptamer no. 20. Lanes:
C, reaction control of the F
N
fraction; F
N

,
RNA fraction which is not bound to Co
2
-
NTA resin; F
B
, RNA fraction that is effec-
tively bound to Co
2+
-NTA resin and is
eluted with a 2 m
M concentration of Co
2+
;
L, formamide ladder; T1, limited hydrolysis
by RNase T
1
. Guanine residues are labeled
on the right. Sites of interference are
denoted with arrows.
Recognition of Co
2+
by RNA aptamers J. Wrzesinski and S. K. Jo
´
z
´
wiakowski
1654 FEBS Journal 275 (2008) 1651–1662 ª 2008 The Authors Journal compilation ª 2008 FEBS
no. 20, respectively. Additional weak interferences
(j < 2) at G39 and G40 were discovered. According

to our earlier assumption, in aptamer no. 20, the self-
complementary region 35-ACGCGG-40 was predicted
to be involved in the formation of the kissing dimer
[19]. The appearance of this effect in aptamer no. 20
was confirmed experimentally (Fig. 6). In the presence
of a 1 mm concentration of Co
2+
, as well as Mg
2+
ions, mobility shift on the nondenaturing gel was
observed, indicating that divalent metal ions Co
2+
and
Mg
2+
are necessary for kissing dimer formation in a
nonspecific manner. We believe that interferences
within the self-complementary region of aptamer
no. 20 indicate the additional stabilization of the kiss-
ing complex by interaction of Co
2+
with the N7 atom
of guanine bases, namely nonstandard G–A and the
neighboring G–C base pairs. Interestingly, analysis of
the crystal structure of RNA fragments mimicking the
HIV-1 virus subtype A kissing complex, crystallized in
the presence of different metals, revealed that Mg
2+
,
Fig. 4. Summary of the interference effects within the Co

2+
bind-
ing aptamer no. 18 and no. 20 defined by NAIM. The histogram
represents the secondary structure of the aptamer loop regions.
The bars are correlated with the determined magnitude of interfer-
ence (j values). Analogs used in NAIM are marked with appropriate
colors. Dotted lines in aptamer no. 18 mark the loop E-like base
pairing, and the solid line in aptamer no. 20 indicates the nucleo-
tides involved in the dimer complex.
A
B
C
Fig. 5. The aptamer no. 18 loop E-like motif structure: (A) nucleo-
tides involved in loop E motif formation; (B) hydrogen bonding pat-
terns in the extended G-A; and (C) reversed Hoogsteen A-U
nonstandard base pairs.
1 mM EDTA
Aptamer no. 18
Aptamer no. 20
dime
r
H
2
O
+
++
++
++
+–
–

–
––– –– –
–– ––
––––
–
–
–––––
1 m
M MgCl
2
1 mM CoCl
2
Fig. 6. (A) Gel shift assay dimer formation by the aptamers no. 18
and no. 20. 5¢-
32
P labelled RNA samples containing 20 mM Tris–HCl
pH 7.5, 40 m
M NaCl dimerization buffer were supplemented with
1m
M concentration of EDTA, Mg
2+
and Co
2+
, respectively. (B)
Scheme of proposed secondary structure of aptamer no. 20 dimer
complex. The self-complementary sequence ACGCGG is shown.
J. Wrzesinski and S. K. Jo
´
z
´

wiakowski Recognition of Co
2+
by RNA aptamers
FEBS Journal 275 (2008) 1651–1662 ª 2008 The Authors Journal compilation ª 2008 FEBS 1655
as well as Co
2+
,Zn
2+
and Mn
2+
, preferentially bind
to the N7 atom of guanine bases in such a motif [25].
Recognition of Co
2+
by RNA aptamers
The data presented here reveal the importance of the
N7 atom of purines in the organization of the Co
2
binding site in selected aptamers and their involvement
in the metal ion coordination (Fig. 4). Participation of
the N7 atom of purines in the organization of the ter-
tiary structure of other RNAs has been extensively
investigated by the NAIM method. Heide et al. [26]
have demonstrated the significance of this position in
purines upon binding of tRNA to Escherichia coli
RNase P. The N7 position of the adenines is also nec-
essary for the catalytic activity of the RNase P–sub-
strate conjugate [27]. Using the NAIM assay, six
adenines critical for self-cleavage have been identified.
The application of a set of phosphorothioate nucleo-

side analogs, including 7-deaza-purine analogs, to
examine the interactions of low molecular ligand glu-
cosamine 6-phosphate with glmS ribozyme in the pres-
ence of Mg
2+
demonstrated the importance of the N7
group of purines and Mg
2+
ions in the organization of
the catalytic site of the ribozyme and in the glucosa-
mine 6-phosphate recognition process [28]. We show
that, in the selected aptamers, the N7 position of some
guanine bases is needed for binding of Co
2+
immobi-
lized on the NTA resin. As Co
2+
ions usually contain
six coordination sites and four of them are occupied
upon complexation with the resin, only two sites
remain available for interactions with RNA ligands,
including the N7 atom of purines. Therefore, there is
the possible involvement of the N7 atom of the neigh-
boring tandem guanines, G27 and G28 or G37 and
G38 in aptamer no. 18, as well as G25 and G26 or
G39 and G40 in aptamer no. 20, in Co
2+
binding.
Interestingly, a similar metal ion binding mode to
the tandem guanines has been observed in the resolved

crystal structure of leadzyme [29]. Strontium ions,
which mimic lead ions, have been bound directly or
water mediated to the N7 position of several purines,
mainly guanine tandems. In other RNAs whose struc-
tures have been determined with atomic resolution,
such as hairpin ribozyme and the P4–P6 domain of the
group I intron, binding of metal ions in a manner
identical to that of guanine tandems has been identi-
fied [30,31].
The previously performed Co
2+
-induced cleavage
of aptamers revealed a doublet of scissions occurring
at nucleotides G37 and G38 in aptamer no. 18 and
nucleotides A31 and G32 in aptamer no. 20 [19].
Additionally, the determined cleavage rate constant for
aptamer no. 18 was three-fold higher than that for apt-
amer no. 20. It is well established that the rate of
metal ion-induced cleavage strongly depends on the
distance between the ion in its strong binding site and
the 2¢OH group of ribose moiety involved in the
scission phosphodiester bond mechanism [10]. Thus,
R
P
-phosphorothioate CaS interference at C36 in apt-
amer no. 18, adjacent to Co
2+
-induced cleavage sites,
strongly suggests that this region is involved in direct
metal ion–RNA interactions. In the case of aptamer

no. 20, those interactions are presumably weaker;
hence, interference was not observed.
Influence of Co
2+
on aptamer structures in
solution
To gain a better insight into the effect of Co
2+
binding
on the RNA structure, we applied a semi-random DNA
library of 6-mers and RNase H, an endonucleolytic
ribonuclease that specifically recognizes the DNA–RNA
duplex and digests it. It has been shown that hybridiza-
tion of short oligodeoxyribonucleotides to RNAs is
strongly affected by RNA target structures, which
results in changing RNase H digestion efficiency [32,33].
One big advantage of this approach involving the appli-
cation of a semi-random library is that no knowledge of
the RNA structure is required to determine the DNA
oligomer sequence that effectively hybridizes to the
RNA target. We applied a semi-random library contain-
ing a single fixed nucleotide (A,G, C or T), located in
the third position of the oligodeoxyribonucleotide chain
and five random nucleotides; thus, the library consisted
of 4096 members (Fig. 7A). Knowing the RNase H
digestion preferences that cleave RNA at the end of
bound DNA oligomer, it is possible to correlate the
RNase H digestion sites with the most likely positions
of hybridized DNA 6-mers [32,33].
In a first step, we determined the RNase H digestion

sites and the possible location of the binding region
within the aptamer structures to which 6-mers, mem-
bers of the oligodeoxyribonucleotide library, hybridize.
In aptamer no. 18, digestions took place at G27, G28,
G30 and G31 at the 5¢-end of the loop (Fig. 7B,C).
Additionally, a doublet at G35 and C36 was identified.
However, in aptamer no. 20, four digestion sites were
found: G26, U27, A30 and A31 at the 5¢-end of loop.
In both loops, 6-mer oligomers hybridize to the 5¢-end
of the loops and propagate to nucleotides involved in
the formation of the helix. The strong preference for
the 5 ¢-side of the loops in comparison with the 3 ¢ -side
may be explained by different sequences of both
sides. As noted above, both aptamers contain purine
rich stretches that could be involved in stacking
Recognition of Co
2+
by RNA aptamers J. Wrzesinski and S. K. Jo
´
z
´
wiakowski
1656 FEBS Journal 275 (2008) 1651–1662 ª 2008 The Authors Journal compilation ª 2008 FEBS
interactions. Such specific characteristics of RNA,
involving the U turn, the formation of stable tertiary
base pairs and, particularly, the stacking interactions
that ensure the helical order of single-stranded regions,
are the main factors determining the efficiency of
hybridization of short DNA oligomers to the RNA
target [32,33]. Protection of the 35-ACG-37 region in

aptamer no. 20 against hybridization of the 6-mers oli-
godeoxyribonucleotide library is due to the formation
of the kissing loop complex (Fig. 6). Additionally, the
observations of RNase H specificity are well correlated
with the results obtained by NAIM noted earlier. The
main c
7
GaS interference sites are positioned in the
same regions: 19-AGGCGAGAGG-28 at G27 and
A
B
C
Fig. 7. (A) Sequence of the oligodeoxyribo-
nucleotide library used. (B) Autoradiograms
of RNA fragments showing digestion of
Co
2+
binding aptamers with RNase H in the
presence of semi-random libraries. 5¢-
32
P
end-labeled RNAs were used and the reac-
tion products were analyzed on the gel.
Lanes: –, reaction control; a, g, c, t, parts of
semi-random library; L, formamide ladder;
T1, limited hydrolysis by RNase T
1
. Guanine
residues are labeled on the right. (C) Loca-
tion of RNase H digestion sites displayed on

the aptamer secondary structure models.
Gray lines along the aptamer sequences
show the possible position of 6-mer oligode-
oxyribonucleotides hybridizing to the RNA
targets. Letters with arrow denote diges-
tions to which the corresponding 6-mer oli-
godeoxyribonucleotides could be assigned.
J. Wrzesinski and S. K. Jo
´
z
´
wiakowski Recognition of Co
2+
by RNA aptamers
FEBS Journal 275 (2008) 1651–1662 ª 2008 The Authors Journal compilation ª 2008 FEBS 1657
G28 in aptamer no. 18 and 19-AGGCGAGG-26 at
G25 and G26 in aptamer no. 20, which we postulate
to be involved in Co
2+
coordination.
The dependence of oligomer hybridization and
RNase H digestion of aptamers on Co
2+
concentration
was studied with a library (i.e. the part of the semi-
random library that contains adenosine in the third
position) because this part of the library revealed the
most prominent and specific RNase H digestion sites
(Fig. 7B,C). We detected inhibition of RNase H diges-
tion efficiency with an increased Co

2+
concentration
despite the presence of Mg
2+
at a 10 mm concentra-
tion, as necessary for enzyme activity (Fig. 8). The
inhibition constants, determined as the concentration
of Co
2+
at which the extent of RNase H digestion was
reduced by half, were approximately 0.5 mm for both
aptamers. Subsequently, to exclude the possibility that
Co
2+
would inhibit RNase H activity at a higher con-
centration, we used another RNA model, antigenomic
delta ribozyme, which has been well characterized in
our laboratory [34]. Previously, we mapped the hybrid-
ization sites of the 6-mer oligodeoxyribonucleotide
library in delta ribozyme using an RNase H digestion
assay and they have appeared to be localized within
the single-stranded P1 and J1 ⁄ 4 regions [32]. Delta
ribozyme is highly active in the presence of a 1 mm
concentration of Co
2+
ions; therefore, these ion bind-
ing sites probably occur inside the ribozyme structure
[35]. Strikingly, we did not observe reduction of the
RNase H digestion efficiency, but an increase of the
digestion yield when Co

2+
was added, even at a 3 mm
concentration. We assume that the lower RNase H
digestion efficiency of the studied RNA aptamers is
mainly related to the stabilization of their structures in
the presence of Co
2+
ions. Presumably, the regions to
which DNA oligomers hybridize become less accessible
in more compact aptamer structures upon Co
2+
bind-
ing and this process prevents the DNA–RNA duplex
formation necessary for the RNase H digestion event.
A reverse effect takes place in the antigenomic delta
ribozyme. The presence of Co
2+
presumably desta-
bilizes or rearranges the ribozyme structure facilitating
hybridization of DNA oligomers because an increase
of RNase H digestion efficiency was observed.
The influence of Co
2+
ions on the global structure of
aptamers has been studied by applying the UV melting
technique (data not shown). We observed an increase
of the melting temperature from 69.7 °C to 71.8 °C for
aptamer no. 18, and from 64.9 °C to 66.2 °C for
aptamer no. 20, despite the low concentration of Co
2+

0
0.01 0.06
Aptamer no. 18 Aptamer no. 20 Antigenomic delta ribozyme
0.1 0.2 0.3 0.5 1 2 30.03 L T1
G38
G35
G31
C
0
0.01 0.06
0.1 0.2 0.3 0.5 1 2 30.03 L T1
C
0 0.1 0.2 0.3 0.5 1 2 30.05 L T1
CC1
Co [mM]
2+
G32
G37
G29
G26
G42
G35
0.0 1.0 2.0 3.00.5 1.5 2.5
Co
2+
concentration [mM]
0.0 1.0 2.0 3.00.5 1.5 2.5
Co
2+
concentration [mM]

0.0 1.0 2.0 3.00.5 1.5 2.5
Co
2+
concentration [mM]
0.0
0.2
0.4
Fraction cleaved
0.6
0.8
1.0
0.0
0.2
0.4
Fraction cleaved
0.6
0.8
1.0
0.0
0.2
0.4
Fraction cleaved
0.6
0.8
1.0
A
B
Fig. 8. (A) Digestion of Co
2+
binding aptamers and antigenomic delta ribozyme with RNase H in the presence of a-library (ie: the part of the

semi-random library that contains adenosine in the third position) at different Co
2+
concentrations. Figure labelling is the same as in Fig. 7.
C, incubation control; C1, incubation control in the presence of a 3 m
M Co
2+
concentration. (B) Graphical representation of the dependencies
of RNase H digestion efficiency on Co
2+
concentration of aptamers no. 18 and no. 20 and antigenomic delta ribozyme.
Recognition of Co
2+
by RNA aptamers J. Wrzesinski and S. K. Jo
´
z
´
wiakowski
1658 FEBS Journal 275 (2008) 1651–1662 ª 2008 The Authors Journal compilation ª 2008 FEBS
(0.1 mm). At a higher Co
2+
concentration, significant
degradation of RNA during UV melting experiments
was detected. This observation supports the above
suggestion, as formulated on the basis of oligomer
hybridization and RNase H digestion, that the binding
of Co
2+
to in vitro selected Co
2+
-specific aptamers

stabilizes their structure, even at a low metal ion
concentration.
In the present study, we postulate that Co
2+
binding
to the aptamer structures induces the increase in their
stabilities. A similar situation has been observed for
riboswitches, which are highly structured domains that
reside in the 5¢-UTR region of mRNA and affect gene
expression [36]. The binding of some low molecular
ligands (e.g. FMN, adenosylocobalamin, guanine,
l-lysine) to the ‘aptamer domain’ of riboswitches stabi-
lizes the specific RNA structure involved in the regula-
tion of gene expression at a transcription or translation
level. It is also known that attachment of the Co
2+
binding domains to the allosteric hammerhead ribo-
zyme may regulate its catalytic activity [37]. More
recently, riboswitches that respond to the cellular con-
centration of Mg
2+
by forming a stable hairpin that
affects the transcription level of Mg
2+
transporter gene
expression (MgtA and MgtE in bacteria Salmonella
enterica and Bacillus subtilis, respectively) have been
described [38,39]. The above information, together with
the data presented here, indicates that possibly also
Co

2+
binding RNA molecules responding to the Co
2+
concentration may regulate RNA activity.
Conclusions
The significance of Mg
2+
for RNA stability, RNA
catalysis and the regulation of gene expression is well
established. However, information concerning partici-
pation of other metal ions with distinct chemical prop-
erties (e.g. Co
2+
, a member of the transition metal ion
group), in such processes, and particularly the molecu-
lar recognition of Co
2+
by RNA, still remains limited
and requires further study.
The application of phosphorothioate 7-deaza-purine
analogs and the NAIM approach for studying the
structure of Co
2+
binding sites in aptamers has
revealed the importance of the N7 group of guanine
bases. In both aptamers, we identified several tandems
of guanines involved in Co
2+
binding or the develop-
ment of aptamer structures. Additionally, other ele-

ments of the secondary structure of aptamers, such as
the E-like loop and kissing complex motifs, appear to
be important in the formation of the architecture of
Co
2+
binding sites. However, in the case of aptamer
no. 20, which contains the self-complementary 35-AC
GCGG-40 sequence, we have confirmed the appear-
ance of a kissing loop complex motif in the presence
of divalents Co
2+
or Mg
2+
. This observation is in line
with NAIM results showing that the N7 groups of
G39 and G40 interfere with Co
2+
.
We would like to emphasize the importance of the
purine rich stretches with stacking interactions for the
binding of Co
2+
and other nucleic acid molecules.
Regions 19-AGGCGAGAGG-28 in aptamer no. 18
and 19-AGGCGAGG-26 in aptamer no. 20 contain a
tandem of guanines with the highest c
7
GaS interfer-
ence values, thus indicating direct coordination of
Co

2+
. The same regions are targets for 6-mer oligode-
oxyribonucleotides, as confirmed in the present study
using semi-random DNA library hybridization and an
RNase H digestion assay.
Moreover, the binding of Co
2+
to aptamers induces
conformational changes that result in the stabilization
of the RNA structures, which was confirmed in two
independent experiments, namely (a) the hybridization
of a semi-random library and RNase H digestion and
(b) temperature-dependent UV melting.
Experimental procedures
Materials
NTA resin was obtained from Novagen (Darmstadt, Ger-
many); all chemicals were obtained from Serva (Heidelberg,
Germany) or Fluka (Buchs, Switzerland). Phosphorothioate
nucleotides (NTPaS) and c
7
AaS were purchased from Glen
Research (Starling, VA, USA), except for c
7
GaS, which
was purchased from IBA (Berlin, Germany). Enzymes: T7
RNA polymerase and DNA Taq polymerase T4 polynucleo-
tide kinase were obtained from MBI Fermentas (Vilnius,
Lithuania). [c-
32
P]ATP (5000 CiÆmmol

)1
) was obtained from
Hartmann Analytic (Braunschweig, Germany).
DNA template construction
The following oligodeoxynucleotides were used for con-
struction of the DNA templates: LM47: 5¢-GCGAGCTCT
AATACGACTCACTAT
GGGCATA nCGTTAGGCTGTA
GGC-3¢, LM18: 5¢-CGAAGCTTGCATATGCTACGCT
GAGGCGATAT TTCC GCT TTCC TCTC
GCCTACAGCC
TAACGTATGCCC-3¢ and LM20: 5¢-CGAAGCTTGCA
TATGCTACGCTGAGGCUATTACCGCGTTTCTTCCA
CCTC
GCCTACAGCCTAACGTATGCCC-3¢ (letters in
italic indicate the T7 RNA promotor, complementary
sequences are underlined). Oligomers were deprotected after
synthesis and purified on denaturing 8% (w ⁄ v) polyacryl-
amide gel. Equimolar amounts of oligomers LM 47 and
LM 18 or LM 20 were annealed and double-stranded DNA
J. Wrzesinski and S. K. Jo
´
z
´
wiakowski Recognition of Co
2+
by RNA aptamers
FEBS Journal 275 (2008) 1651–1662 ª 2008 The Authors Journal compilation ª 2008 FEBS 1659
template was generated by PCR. The reaction mixture con-
tained 1.5 lm of both LM 47 oligomers and 0.3 lm of LM

18 or LM20, 10 mm Tris–HCl, pH 7.0, 2 mm MgCl
2
,
150 mm KCl, 0.1% Triton X-100, 200 lm each of dNTP
and 60 UÆmL
)1
of DNA Taq polymerase. The reaction was
performed on Biometra (Go
¨
ettingen, Germany) UNO II
thermocycler for six cycles (30 s at 93 °C, 30 s at 55 °C and
1 min at 72 °C). The double-stranded DNA was extracted
with phenol ⁄ chloroform (1 : 1), precipitated with buffer:
3 m sodium acetate, pH 7.0 ⁄ ethanol (1 : 9) at )20 °C over-
night. The dsDNA template was recovered by centrifuga-
tion, dissolved in TE buffer and used in the transcription
reaction.
RNA preparation
The modified RNA aptamers were prepared by the in vitro
transcription reaction where the typical transcription reac-
tion contained 0.5 lm dsDNA template, 40 mm Tris–HCl,
pH 7.0, 10 mm MgCl
2
, Triton X-100, 2 mm spermidine,
5mm dithiothreitol, 1 mm each NTP, and 2000 UÆmL
)1
T7
RNA polymerase. Additionally, the reaction mixtures con-
tained one from the following phosphorothioate analogs:
0.05 mm ATPaS, 0.07 mm UTPaS, 0.07 mm CTPaS,

0.1 mm GTPaS, as well as 0.05 mm c
7
ATPaS, 0.1 mm
c
7
GTPaS. To facilitate 5¢ labeling, 4 mm guanosine was
added. Following incubation of the mixture at 37 °C for
4 h and purification on polyacrylamide gel, the RNA tran-
scripts were excised, eluted, precipitated with ethanol, and
RNA was recovered by centrifugation and dissolved in
10 mm Tris–HCl, pH 7.0, 0.1 mm EDTA.
Nucleotide analogs interference mapping
Prior to the interference procedure, RNA transcripts were
5¢-end labeled with [c-
32
P]ATP and polynucleotide kinase
under standard conditions. Labeled aptamers (typically
2–3 · 10
6
c.p.m., 20 pmol of RNA supplemented with
unlabeled RNA to a final 0.2 lm concentration) were sub-
jected to a denaturation–renaturation procedure in 200 lL
of the standard binding buffer A (40 mm Hepes, pH 8.0,
400 mm NaCl, 1 mm MgCl
2
); RNA was incubated at 70 °C
for 5 min, and slowly cooled to 25 °C. The aptamers were
bound to Co
2+
-NTA resin, mixed gently with resin slurry

for 5 min and, subsequently, the resin was washed with six
volumes (400 lL each) of buffer A to remove the unbound
RNA. In the next step, the resin was washed with addi-
tional six volumes of buffer A containing 2 mm CoCl
2
.To
both fractions, 1 ⁄ 10 volume of sodium acetate (pH 7.0) and
3 volumes of ethanol were added and the mixtures were
stored at )20 °C overnight. The RNA was recovered by
centrifugation and dissolved in water. The RNA fraction
containing molecules that specifically bind to Co
2
-NTA
resin (F
B
) and the fraction deprived of such properties (F
N
)
were subjected to cleavage of phosphorothioate linkage in
the presence of 10 mgÆmL
)1
iodine ⁄ ethanol solution. The
cleavage was carried out at 37 °C for 15 min; subsequently,
the RNA was precipitated, recovered by centrifugation and
dissolved in loading buffer (7 m urea ⁄ dyes 10 mm EDTA).
Patterns of I
2
cleavage were compared on 8 m urea ⁄ 12%
polyacrylamide gel alongside alkaline hydrolysis ladders
and partial digestion with RNase T

1
. Products were
visualized by autoradiography or quantified using a
PhosphoImager Typhoon 8600 (Uppsala, Sweden) with
imagequant software (Uppsala, Sweden).
The interference j value was calculated from the equation:
j ¼
band intensity at nucleotide x
ðÞ
P
band intensity at nucleotide
aÀz
ðÞ
NTP
(band intensity at nucleotide xÞ
P
band intensity at nucleotide
aÀz
ðÞ
NTPaS
Interferences were considered as: weak, j = 1.5–2.0; mod-
erate, 2.0–2.5; and strong, > 2.5.
RNase H RNA mapping experiment
The RNA aptamers were prepared by in vitro transcription,
under the conditions described above, using unmodified
NTPs, purified and labeled at their 5¢-end. Prior to digestion
with E. coli, RNase H 5¢-
32
P labeled RNA was renatured in
buffer (40 mm Tris–HCl, pH 8.0, 40 mm KCl, 10 mm MgCl

2
,
1mm dithiothreitol and 0.1 mm EDTA). Subsequently,
RNase H was added to a final concentration of 250 UÆmL
)1
.
The digestion reactions were initiated by adding separately
four parts of the DNA 6-mers library with the appropriate
fixed nucleotide to four RNA target samples to a final con-
centration of 200 lm (i.e. 5000-fold excess over the RNA
concentration). The mixtures were incubated at 37 °C for
30 min, quenched with an equal volume of 7 m urea ⁄ 20 mm
EDTA and immediately frozen on dry ice.
In vitro RNA aptamer dimerization assay
Aliquots of 5¢ labeled RNA aptamers, supplemented with
unlabeled RNA to a concentration of 250 nm, were heated
at 100 °C for 1 min and cooled on ice for 10 min. Then
dimerization buffer containing, respectively, 20 mm Tris-
HCl, pH 7.5, 40 mm NaCl alone, or supplemented with
1mm EDTA, MgCl
2
and CoCl
2
, was added. Subsequently,
the reaction mixture was heated for 10 min at 25 °C. After
adding 1 ⁄ 5 volume 30% glycerol, the samples were loaded
directly on the 12% nondenaturing polyacrylamide gel. The
electrophoresis under nondenaturing conditions was carried
out at room temperature using 20 mm Tris–HCl, pH 7.5,
40 mm NaCl, 1 mm MgCl

2
electrophoresis buffer.
Acknowledgements
We are very grateful to Professor Jerzy Ciesiolka for
critically reading the manuscript and for helpful
Recognition of Co
2+
by RNA aptamers J. Wrzesinski and S. K. Jo
´
z
´
wiakowski
1660 FEBS Journal 275 (2008) 1651–1662 ª 2008 The Authors Journal compilation ª 2008 FEBS
discussions throughout the course of this work. We
wish to thank Ms Barbara Smo
´
lska for her excellent
technical assistance. We also thank the reviewers for
their comments on the manuscript. This work was sup-
ported by grant 6 P04A 081 21, from the Polish Com-
mittee for Scientific Research as well as the Bioorganic
Chemistry and Structural Biology Network.
References
1 Mello CC & Conte DJ (2004) Revealing the world of
RNA interference. Nature 431, 338–342.
2 Filipowicz W, Jaskiewicz L, Kolb FA & Pillai RS
(2006) Post-transcriptional gene silencing by siRNAs
and miRNAs. Curr Opin Struct Biol 15 , 331–341.
3 Szymanski M, Barciszewska MZ, Erdmann V & Bar-
ciszewski J (2005) A new frontier for molecular medi-

cine: noncoding RNAs. Biochim Biophys Acta 1756,
65–75.
4 Pyle A (2002) Metal ions in the structure and function
of RNA. J Biol Inorg Chem 7, 679–690.
5 Draper DE, Grilley D & Soto AM (2005) Ions and
RNA folding. Annu Rev Biophys Biomol Struc 34,
21–243.
6 Egli M (2004) Nucleic acid crystallography: current pro-
gress. Curr Opin Chem Biol 8, 580–591.
7 Furtig B, Richter C, Wohnert J & Schwalbe H (2003)
NMR spectroscopy of RNA. Chembiochem 4, 936–962.
8 Michalowski D, Wrzesinski J & Krzy
_
zosiak WJ (1996)
Cleavages induced by different metal ions in yeast
tRNA(Phe) U59C60 mutants. Biochemistry 35, 10727–
10734.
9 Matysiak M, Wrzesinski J & Ciesiolka J (1999) Sequen-
tial folding of the genomic ribozyme of the hepatitis
delta virus: structural analysis of RNA transcription
intermediates. J Mol Biol 291, 283–294.
10 Kirsebom LA & Ciesiolka J (2005) Pb
2+
-induced
cleavage of RNA. In Handbook of RNA Biochemistry
(Hartmann RK, Bindereif A, Schon A & Westhof E,
eds), pp. 214–228. Wiley-VCH GmbH & Co, KGaA,
Weinheim.
11 Pan T, Long DM & Uhlenbeck OC (1993) Divalent
metal ions in RNA folding and catalysis. In RNA

World (Gesteland RF & Atkins JF, eds), pp. 271–302.
Cold Spring Harbor Laboratory Press, Plainview, NY.
12 Ruffner DE & Uhlenbeck OC (1990) Thiophosphate
interference experiments locate phosphates important
for the hammerhead RNA self-cleavage reaction.
Nucleic Acids Res 18, 6025–6029.
13 Basu S & Strobel SA (1999) Thiophilic metal ion rescue
of phosphorothioate interference within the Tetrahy-
mena ribozyme P4-P6 domain. RNA 5, 1399–1407.
14 Ryder SP & Strobel SA (1999) Nucleotide analog inter-
ference mapping. Methods 18, 38–50.
15 Strobel SA (1998) Ribozyme chemogenetics. Biopoly-
mers 48, 65–81.
16 Kuzic S & Hartmann RK (2005) Nucleotide analog
interference mapping in the RNase P system. In Hand-
book of RNA Biochemistry (Hartmann RK, Bindereif
A, Schon A & Westhof E, eds), pp. 294–318. Wiley-
VCH GmbH & Co, KGaA, Weinheim.
17 Szewczak AA, Kosek A B, Piccirilli JA & Strobel SA (2002)
Identification of an active site ligand for a group I ribo-
zyme catalytic metal ion. Biochemistry 41, 2516–2525.
18 Ryder SP & Strobel SA (1999) Nucleotide analog inter-
ference mapping of the hairpin ribozyme: implications
for secondary and tertiary structure formation. Nucleic
Acids Res 30, 1287–1291.
19 Wrzesinski J & Ciesiolka J (2005) Characterization of
structure and metal ions specificity of Co
2+
-binding
RNA aptamers. Biochemistry 44, 6257–6268.

20 Feig AL & Uhlenbeck OC (1999) The role of metal ions
in RNA biochemistry. In RNA World (Gesteland RF,
Cech TR & Atkins G eds), pp. 287–319. Cold Spring
Harbor Laboratory Press, Plainview, NY.
21 Rich A & RajBhandary UL (1976) Transfer RNA:
molecular structure, sequence, and properties. Annu Rev
Biochem 45, 805–860.
22 Adams PL, Stahley MR, Gill ML, Kosek AB, Wang J
& Strobel SA (2005) Crystal structure of a group I
intron splicing intermediate. RNA 10, 1867–1887.
23 Kim SH, Shin WC & Warrant RW (1985) Heavy metal
ion-nucleic acid interaction. Methods Enzymol 114,
156–167.
24 Leontis NB & Westhof E (1998) The 5S rRNA loop E:
chemical probing and phylogenetic data versus crystal
structure. RNA 4, 1134–1153.
25 Ennifar E, Walter P & Dumas P (2003) A crystallo-
graphic study of the binding of 13 metal ions to two
related RNA duplexes. Nucleic Acids Res 31, 2671–2682.
26 Heide C, Feltens R & Hartmann RK (2001) Purine N7
groups that are crucial to the interaction of Escherichia
coli RNase P RNA with tRNA. RNA 7, 958–968.
27 Kazantsev AV & Pace NR (1998) Identification by
modification-interference of purine N-7 and ribose 2¢-
OH groups critical for catalysis by bacterial ribonucle-
ase P. RNA 4, 937–947.
28 Jansen JA, McCarthy TJ, Soukup GA & Soukup JK
(2006) Backbone and nucleobase contacts to glucosa-
mine-6-phosphate in the glmS ribozyme. Nat Struct
Mol Biol 13, 517–523.

29 Wedekind JE & McKay DB (2003) Crystal structure of
the leadzyme at 1.8 A resolution: metal ion binding and
the implications for catalytic mechanism and allo site
ion regulation. Biochemistry 42, 9554–9563.
30 Butcher SE, Allain FH & Feigon J (2000) Determina-
tion of metal ion binding sites within the hairpin ribo-
zyme domains by NMR. Biochemistry 7, 2174–2182.
J. Wrzesinski and S. K. Jo
´
z
´
wiakowski Recognition of Co
2+
by RNA aptamers
FEBS Journal 275 (2008) 1651–1662 ª 2008 The Authors Journal compilation ª 2008 FEBS 1661
31 Cate JH & Doudna JA (1996) Metal-binding sites in
the major groove of a large ribozyme domain. Structure
4, 1221–1229.
32 Wrzesinski J, Legiewicz M & Ciesiolka J (2000) Map-
ping of accessible sites for oligonucleotide hybridization
on hepatitis delta virus ribozymes. Nucleic Acids Res 28,
1785–1793.
33 Mir KU & Southern EM (1999) Determining the influ-
ence of structure on hybridization using oligonucleotide
arrays. Nat Biotechnol 17, 788–792.
34 Wrzesinski J, Legiewicz M, Smolska B & Ciesiolka J
(2001) Catalytic cleavage of cis- and trans-acting antige-
nomic delta ribozymes in the presence of various diva-
lent metal ions. Nucleic Acids Res 29, 4482–4492.
35 Ke A, Zhou K, Ding F, Cate JH & Doudna JA (2004)

A conformational switch controls hepatitis delta virus
ribozyme catalysis. Nature 429, 201–205.
36 Winkler WC & Breaker RR (2005) Regulation of bacte-
rial gene expression by riboswitches. Annu Rev Micro-
biol 59, 487–517.
37 Zivarts M, Liu Y & Breaker RR (2005) Engineered
allosteric ribozymes that respond to specific divalent
metal ions. Nucleic Acids Res 33, 622–631.
38 Cromie JM, Shi Y, Latiff T & Groisman EA (2006) An
RNA sensor for Intracellular Mg
2+
. Cell 125, 71–84.
39 Dann CE III, Wakeman CA, Sieling CL, Baker SC,
Imov I & Winkler WC (2007) Structure and mechanism
of a metal-sensing regulatory RNA. Cell 130, 878–892.
Recognition of Co
2+
by RNA aptamers J. Wrzesinski and S. K. Jo
´
z
´
wiakowski
1662 FEBS Journal 275 (2008) 1651–1662 ª 2008 The Authors Journal compilation ª 2008 FEBS